Autism is a pervasive developmental disorder involving significant impairments in social and communicative functioning and restricted activities and interested. Prospective longitudinal studies are uncommon and, yet, are critical to understanding the developmental course of the disorders. We proposed to continue a prospective, longitudinal study of the developing phenotype of autism from toddlerhood to middle childhood that: (1) begins at the earliest age at which diagnosis is possible, (2) uses standardized measures for diagnosis, (3) provides multiple measures of specific developmental processes over time and (4) incorporates several clearly defined control groups. The proposed study is a continuation of an existing study that includes 135 toddlers (32 with autism, 25 with Down syndrome, 28 with fragile x syndrome, 25 with mixed etiology developmental delay, and 25 typically-developing children) There are three waves of data collection (ages 2, 4, and 8) with the goal of examining how various developmental processes interact and influence symptom development. Key areas of interest include: cognitive and language functioning, symptom presentation, affective functioning, intersubjectivity, imitation and praxis, executive function, temperament, problem behavior, and sensory reactivity. Hypothesis regarding rate (i.e., the timing of the emergence of a specific deficit within a specific group), structure (i.e., the characteristic developmental progression of specific skills in a specific group), structure (i.e., the influences of specific details on the development of other skills), and developmental precursors (i.e., determining whether specific skills are necessary and/or sufficient for the development of other skills) will be tested. Analyses will be conducted on both intra-individual and inter-individual level, in order to examine universality, specificity, persistence, and precedence of each domain.